Let’s define selfish and symmetric as those that are within 5 percentage points of being completely selfish or completely symmetric on average. I also look at limits 0.1 and 0.025.

SELFISH_SHARE <- 0.95
SYMMETRIC_SHARE <- 0.05
classifications <- mmzame_decisions %>% mutate(outcome=y/(y+x)) %>%
  group_by(id, treatment) %>% 
  summarize(mean_yshare = mean(outcome)) %>%
  spread(treatment, mean_yshare) %>%
  mutate(selfish = (dictator>=SELFISH_SHARE),
         symmetric = abs(dictator - 0.5)<SYMMETRIC_SHARE,
         selfish10 = (dictator>0.9),
         symmetric10 = abs(dictator - 0.5)<0.1,
         selfish025 = (dictator>0.975),
         symmetric025 = abs(dictator - 0.5)<0.025) 
classifications %>% ungroup %>% summarise_all(funs(mean)) %>%
  dplyr::select(selfish:symmetric025) 
## Warning: funs() is soft deprecated as of dplyr 0.8.0
## Please use a list of either functions or lambdas: 
## 
##   # Simple named list: 
##   list(mean = mean, median = median)
## 
##   # Auto named with `tibble::lst()`: 
##   tibble::lst(mean, median)
## 
##   # Using lambdas
##   list(~ mean(., trim = .2), ~ median(., na.rm = TRUE))
## This warning is displayed once per session.
## # A tibble: 1 x 6
##   selfish symmetric selfish10 symmetric10 selfish025 symmetric025
##     <dbl>     <dbl>     <dbl>       <dbl>      <dbl>        <dbl>
## 1   0.373    0.0688     0.467       0.123      0.312       0.0399
classifications <- classifications %>%
  dplyr::select(id, selfish, symmetric)

Scatter graphs

ids <- unique(mmzame_decisions$id)
captions <- list()
selfishgraphs <- list()
for (id in ids) {
  if (classifications$selfish[classifications$id==id] == TRUE) {
    caption <- "selfish"
  } else if (classifications$symmetric[classifications$id==id]==TRUE) {
    caption <- "impartial"
  } else {
    caption <- "neither"
  }
  captions[[id]] <- caption
}
scatterdata <- mmzame_decisions %>% 
  mutate(logP = log(maxx/maxy),
         treatment = fct_recode(as_factor(treatment),
                                "Social Choice" = "dictator",
                                "Social Risk" = "moral",
                                "Personal Risk" = "risk"))
for (id0 in ids) {
  caption <- captions[[id0]]
  if (caption == "selfish") {
    fname <- paste0("logpricegraphs/selfish/yshare_logp_",id0,".pdf")
    caption <- paste0("This individual, id=", id0,", is selfish.")
    fname2 <- paste0("logpricegraphs/selfish/yshare_logp_selfish_implicit_",id0,".pdf")
    sg <- scatterdata %>% filter(id==id0) %>%
      filter(treatment!="Social Choice") %>%
      ggplot(aes(x=logP, y=yshare, shape=treatment)) +
      geom_point(size=2) + 
      scale_shape_manual(values=c(1,3)) +
      labs(x = "log(p)", 
           y = "y/(x+y)",
           caption = caption) +
      theme_bw() + theme(legend.position = "bottom", 
                         legend.justification = "center",
                         legend.direction = "horizontal",
                         legend.box = "horizontal")
    selfishgraphs[[id0]] <- sg
    print(sg)
    ggsave(here("graphs", fname2), width=7,height=5)
  } else if (caption == "impartial") {
    fname <- paste0("logpricegraphs/impartial/yshare_logp_",id0,".pdf")
    caption <- "This individual is impartial."
  } else {
    fname <- paste0("logpricegraphs/neither/yshare_logp_",id0,".pdf")
    caption <- "This individual is neither selfish nor impartial."
  }
  scatterdata %>% filter(id==id0) %>%
  ggplot(aes(x=logP, y=yshare, color=treatment)) + 
    geom_point(position="jitter") +
    labs(x = "log(p)",
         y = "y/(x+y)",
         caption = caption) +
    theme_bw() + theme(legend.position = c(0.88,0.8))
  ggsave(here("graphs", fname), width=7, height=5 )
}

Now, we want to select 4 of these for a graph in the paper.

p511 <- selfishgraphs[[511]] + theme(legend.position='none') + labs(x="log(p), id=511", caption="")
p635 <- selfishgraphs[[635]] + theme(legend.position='none') + labs(x="log(p), id=635", caption="")
p317 <- selfishgraphs[[317]] + theme(legend.position='none') + labs(x="log(p), id=317", caption="")
p645 <- selfishgraphs[[645]] + theme(legend.position='none') + labs(x="log(p), id=645", caption="") 

p511 + p635 + p317 + p645 +
  plot_annotation(tag_levels = 'A') + theme(legend.position="bottom", legend.justification = "left")

ggsave(here("graphs", "logprice_scatters.pdf"))
## Saving 7 x 5 in image